unittests/Support/JSONTest.cpp

   1 //===-- JSONTest.cpp - JSON unit tests --------------------------*- C++ -*-===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #include "llvm/Support/JSON.h"
  10 #include "llvm/Support/raw_ostream.h"
  11
  12 #include "gmock/gmock.h"
  13 #include "gtest/gtest.h"
  14
  15 namespace llvm {
  16 namespace json {
  17
  18 namespace {
  19
  20 std::string s(const Value &E) { return llvm::formatv("{0}", E).str(); }
  21 std::string sp(const Value &E) { return llvm::formatv("{0:2}", E).str(); }
  22
  23 TEST(JSONTest, Types) {
  24   EXPECT_EQ("true", s(true));
  25   EXPECT_EQ("null", s(nullptr));
  26   EXPECT_EQ("2.5", s(2.5));
  27   EXPECT_EQ(R"("foo")", s("foo"));
  28   EXPECT_EQ("[1,2,3]", s({1, 2, 3}));
  29   EXPECT_EQ(R"({"x":10,"y":20})", s(Object{{"x", 10}, {"y", 20}}));
  30
  31 #ifdef NDEBUG
  32   EXPECT_EQ(R"("��")", s("\xC0\x80"));
  33   EXPECT_EQ(R"({"��":0})", s(Object{{"\xC0\x80", 0}}));
  34 #else
  35   EXPECT_DEATH(s("\xC0\x80"), "Invalid UTF-8");
  36   EXPECT_DEATH(s(Object{{"\xC0\x80", 0}}), "Invalid UTF-8");
  37 #endif
  38 }
  39
  40 TEST(JSONTest, Constructors) {
  41   // Lots of edge cases around empty and singleton init lists.
  42   EXPECT_EQ("[[[3]]]", s({{{3}}}));
  43   EXPECT_EQ("[[[]]]", s({{{}}}));
  44   EXPECT_EQ("[[{}]]", s({{Object{}}}));
  45   EXPECT_EQ(R"({"A":{"B":{}}})", s(Object{{"A", Object{{"B", Object{}}}}}));
  46   EXPECT_EQ(R"({"A":{"B":{"X":"Y"}}})",
  47             s(Object{{"A", Object{{"B", Object{{"X", "Y"}}}}}}));
  48   EXPECT_EQ("null", s(llvm::Optional<double>()));
  49   EXPECT_EQ("2.5", s(llvm::Optional<double>(2.5)));
  50   EXPECT_EQ("[[2.5,null]]", s(std::vector<std::vector<llvm::Optional<double>>>{
  51                                  {2.5, llvm::None}}));
  52 }
  53
  54 TEST(JSONTest, StringOwnership) {
  55   char X[] = "Hello";
  56   Value Alias = static_cast<const char *>(X);
  57   X[1] = 'a';
  58   EXPECT_EQ(R"("Hallo")", s(Alias));
  59
  60   std::string Y = "Hello";
  61   Value Copy = Y;
  62   Y[1] = 'a';
  63   EXPECT_EQ(R"("Hello")", s(Copy));
  64 }
  65
  66 TEST(JSONTest, CanonicalOutput) {
  67   // Objects are sorted (but arrays aren't)!
  68   EXPECT_EQ(R"({"a":1,"b":2,"c":3})", s(Object{{"a", 1}, {"c", 3}, {"b", 2}}));
  69   EXPECT_EQ(R"(["a","c","b"])", s({"a", "c", "b"}));
  70   EXPECT_EQ("3", s(3.0));
  71 }
  72
  73 TEST(JSONTest, Escaping) {
  74   std::string test = {
  75       0,                    // Strings may contain nulls.
  76       '\b',   '\f',         // Have mnemonics, but we escape numerically.
  77       '\r',   '\n',   '\t', // Escaped with mnemonics.
  78       'S',    '\"',   '\\', // Printable ASCII characters.
  79       '\x7f',               // Delete is not escaped.
  80       '\xce', '\x94',       // Non-ASCII UTF-8 is not escaped.
  81   };
  82
  83   std::string teststring = R"("\u0000\u0008\u000c\r\n\tS\"\\)"
  84                            "\x7f\xCE\x94\"";
  85
  86   EXPECT_EQ(teststring, s(test));
  87
  88   EXPECT_EQ(R"({"object keys are\nescaped":true})",
  89             s(Object{{"object keys are\nescaped", true}}));
  90 }
  91
  92 TEST(JSONTest, PrettyPrinting) {
  93   const char str[] = R"({
  94   "empty_array": [],
  95   "empty_object": {},
  96   "full_array": [
  97     1,
  98     null
  99   ],
 100   "full_object": {
 101     "nested_array": [
 102       {
 103         "property": "value"
 104       }
 105     ]
 106   }
 107 })";
 108
 109   EXPECT_EQ(str, sp(Object{
 110                      {"empty_object", Object{}},
 111                      {"empty_array", {}},
 112                      {"full_array", {1, nullptr}},
 113                      {"full_object",
 114                       Object{
 115                           {"nested_array",
 116                            {Object{
 117                                {"property", "value"},
 118                            }}},
 119                       }},
 120                  }));
 121 }
 122
 123 TEST(JSONTest, Parse) {
 124   auto Compare = [](llvm::StringRef S, Value Expected) {
 125     if (auto E = parse(S)) {
 126       // Compare both string forms and with operator==, in case we have bugs.
 127       EXPECT_EQ(*E, Expected);
 128       EXPECT_EQ(sp(*E), sp(Expected));
 129     } else {
 130       handleAllErrors(E.takeError(), [S](const llvm::ErrorInfoBase &E) {
 131         FAIL() << "Failed to parse JSON >>> " << S << " <<<: " << E.message();
 132       });
 133     }
 134   };
 135
 136   Compare(R"(true)", true);
 137   Compare(R"(false)", false);
 138   Compare(R"(null)", nullptr);
 139
 140   Compare(R"(42)", 42);
 141   Compare(R"(2.5)", 2.5);
 142   Compare(R"(2e50)", 2e50);
 143   Compare(R"(1.2e3456789)", std::numeric_limits<double>::infinity());
 144
 145   Compare(R"("foo")", "foo");
 146   Compare(R"("\"\\\b\f\n\r\t")", "\"\\\b\f\n\r\t");
 147   Compare(R"("\u0000")", llvm::StringRef("\0", 1));
 148   Compare("\"\x7f\"", "\x7f");
 149   Compare(R"("\ud801\udc37")", u8"\U00010437"); // UTF16 surrogate pair escape.
 150   Compare("\"\xE2\x82\xAC\xF0\x9D\x84\x9E\"", u8"\u20ac\U0001d11e"); // UTF8
 151   Compare(
 152       R"("LoneLeading=\ud801, LoneTrailing=\udc01, LeadingLeadingTrailing=\ud801\ud801\udc37")",
 153       u8"LoneLeading=\ufffd, LoneTrailing=\ufffd, "
 154       u8"LeadingLeadingTrailing=\ufffd\U00010437"); // Invalid unicode.
 155
 156   Compare(R"({"":0,"":0})", Object{{"", 0}});
 157   Compare(R"({"obj":{},"arr":[]})", Object{{"obj", Object{}}, {"arr", {}}});
 158   Compare(R"({"\n":{"\u0000":[[[[]]]]}})",
 159           Object{{"\n", Object{
 160                             {llvm::StringRef("\0", 1), {{{{}}}}},
 161                         }}});
 162   Compare("\r[\n\t] ", {});
 163 }
 164
 165 TEST(JSONTest, ParseErrors) {
 166   auto ExpectErr = [](llvm::StringRef Msg, llvm::StringRef S) {
 167     if (auto E = parse(S)) {
 168       // Compare both string forms and with operator==, in case we have bugs.
 169       FAIL() << "Parsed JSON >>> " << S << " <<< but wanted error: " << Msg;
 170     } else {
 171       handleAllErrors(E.takeError(), [S, Msg](const llvm::ErrorInfoBase &E) {
 172         EXPECT_THAT(E.message(), testing::HasSubstr(Msg)) << S;
 173       });
 174     }
 175   };
 176   ExpectErr("Unexpected EOF", "");
 177   ExpectErr("Unexpected EOF", "[");
 178   ExpectErr("Text after end of document", "[][]");
 179   ExpectErr("Invalid JSON value (false?)", "fuzzy");
 180   ExpectErr("Expected , or ]", "[2?]");
 181   ExpectErr("Expected object key", "{a:2}");
 182   ExpectErr("Expected : after object key", R"({"a",2})");
 183   ExpectErr("Expected , or } after object property", R"({"a":2 "b":3})");
 184   ExpectErr("Invalid JSON value", R"([&%!])");
 185   ExpectErr("Invalid JSON value (number?)", "1e1.0");
 186   ExpectErr("Unterminated string", R"("abc\"def)");
 187   ExpectErr("Control character in string", "\"abc\ndef\"");
 188   ExpectErr("Invalid escape sequence", R"("\030")");
 189   ExpectErr("Invalid \\u escape sequence", R"("\usuck")");
 190   ExpectErr("[3:3, byte=19]", R"({
 191   "valid": 1,
 192   invalid: 2
 193 })");
 194   ExpectErr("Invalid UTF-8 sequence", "\"\xC0\x80\""); // WTF-8 null
 195 }
 196
 197 // Direct tests of isUTF8 and fixUTF8. Internal uses are also tested elsewhere.
 198 TEST(JSONTest, UTF8) {
 199   for (const char *Valid : {
 200            "this is ASCII text",
 201            "thïs tëxt häs BMP chäräctërs",
 202            "𐌶𐌰L𐌾𐍈 C𐍈𐌼𐌴𐍃",
 203        }) {
 204     EXPECT_TRUE(isUTF8(Valid)) << Valid;
 205     EXPECT_EQ(fixUTF8(Valid), Valid);
 206   }
 207   for (auto Invalid : std::vector<std::pair<const char *, const char *>>{
 208            {"lone trailing \x81\x82 bytes", "lone trailing �� bytes"},
 209            {"missing trailing \xD0 bytes", "missing trailing � bytes"},
 210            {"truncated character \xD0", "truncated character �"},
 211            {"not \xC1\x80 the \xE0\x9f\xBF shortest \xF0\x83\x83\x83 encoding",
 212             "not �� the ��� shortest ���� encoding"},
 213            {"too \xF9\x80\x80\x80\x80 long", "too ����� long"},
 214            {"surrogate \xED\xA0\x80 invalid \xF4\x90\x80\x80",
 215             "surrogate ��� invalid ����"}}) {
 216     EXPECT_FALSE(isUTF8(Invalid.first)) << Invalid.first;
 217     EXPECT_EQ(fixUTF8(Invalid.first), Invalid.second);
 218   }
 219 }
 220
 221 TEST(JSONTest, Inspection) {
 222   llvm::Expected<Value> Doc = parse(R"(
 223     {
 224       "null": null,
 225       "boolean": false,
 226       "number": 2.78,
 227       "string": "json",
 228       "array": [null, true, 3.14, "hello", [1,2,3], {"time": "arrow"}],
 229       "object": {"fruit": "banana"}
 230     }
 231   )");
 232   EXPECT_TRUE(!!Doc);
 233
 234   Object *O = Doc->getAsObject();
 235   ASSERT_TRUE(O);
 236
 237   EXPECT_FALSE(O->getNull("missing"));
 238   EXPECT_FALSE(O->getNull("boolean"));
 239   EXPECT_TRUE(O->getNull("null"));
 240
 241   EXPECT_EQ(O->getNumber("number"), llvm::Optional<double>(2.78));
 242   EXPECT_FALSE(O->getInteger("number"));
 243   EXPECT_EQ(O->getString("string"), llvm::Optional<llvm::StringRef>("json"));
 244   ASSERT_FALSE(O->getObject("missing"));
 245   ASSERT_FALSE(O->getObject("array"));
 246   ASSERT_TRUE(O->getObject("object"));
 247   EXPECT_EQ(*O->getObject("object"), (Object{{"fruit", "banana"}}));
 248
 249   Array *A = O->getArray("array");
 250   ASSERT_TRUE(A);
 251   EXPECT_EQ((*A)[1].getAsBoolean(), llvm::Optional<bool>(true));
 252   ASSERT_TRUE((*A)[4].getAsArray());
 253   EXPECT_EQ(*(*A)[4].getAsArray(), (Array{1, 2, 3}));
 254   EXPECT_EQ((*(*A)[4].getAsArray())[1].getAsInteger(),
 255             llvm::Optional<int64_t>(2));
 256   int I = 0;
 257   for (Value &E : *A) {
 258     if (I++ == 5) {
 259       ASSERT_TRUE(E.getAsObject());
 260       EXPECT_EQ(E.getAsObject()->getString("time"),
 261                 llvm::Optional<llvm::StringRef>("arrow"));
 262     } else
 263       EXPECT_FALSE(E.getAsObject());
 264   }
 265 }
 266
 267 // Verify special integer handling - we try to preserve exact int64 values.
 268 TEST(JSONTest, Integers) {
 269   struct {
 270     const char *Desc;
 271     Value Val;
 272     const char *Str;
 273     llvm::Optional<int64_t> AsInt;
 274     llvm::Optional<double> AsNumber;
 275   } TestCases[] = {
 276       {
 277           "Non-integer. Stored as double, not convertible.",
 278           double{1.5},
 279           "1.5",
 280           llvm::None,
 281           1.5,
 282       },
 283
 284       {
 285           "Integer, not exact double. Stored as int64, convertible.",
 286           int64_t{0x4000000000000001},
 287           "4611686018427387905",
 288           int64_t{0x4000000000000001},
 289           double{0x4000000000000000},
 290       },
 291
 292       {
 293           "Negative integer, not exact double. Stored as int64, convertible.",
 294           int64_t{-0x4000000000000001},
 295           "-4611686018427387905",
 296           int64_t{-0x4000000000000001},
 297           double{-0x4000000000000000},
 298       },
 299
 300       {
 301           "Dynamically exact integer. Stored as double, convertible.",
 302           double{0x6000000000000000},
 303           "6.9175290276410819e+18",
 304           int64_t{0x6000000000000000},
 305           double{0x6000000000000000},
 306       },
 307
 308       {
 309           "Dynamically integer, >64 bits. Stored as double, not convertible.",
 310           1.5 * double{0x8000000000000000},
 311           "1.3835058055282164e+19",
 312           llvm::None,
 313           1.5 * double{0x8000000000000000},
 314       },
 315   };
 316   for (const auto &T : TestCases) {
 317     EXPECT_EQ(T.Str, s(T.Val)) << T.Desc;
 318     llvm::Expected<Value> Doc = parse(T.Str);
 319     EXPECT_TRUE(!!Doc) << T.Desc;
 320     EXPECT_EQ(Doc->getAsInteger(), T.AsInt) << T.Desc;
 321     EXPECT_EQ(Doc->getAsNumber(), T.AsNumber) << T.Desc;
 322     EXPECT_EQ(T.Val, *Doc) << T.Desc;
 323     EXPECT_EQ(T.Str, s(*Doc)) << T.Desc;
 324   }
 325 }
 326
 327 // Sample struct with typical JSON-mapping rules.
 328 struct CustomStruct {
 329   CustomStruct() : B(false) {}
 330   CustomStruct(std::string S, llvm::Optional<int> I, bool B)
 331       : S(S), I(I), B(B) {}
 332   std::string S;
 333   llvm::Optional<int> I;
 334   bool B;
 335 };
 336 inline bool operator==(const CustomStruct &L, const CustomStruct &R) {
 337   return L.S == R.S && L.I == R.I && L.B == R.B;
 338 }
 339 inline llvm::raw_ostream &operator<<(llvm::raw_ostream &OS,
 340                                      const CustomStruct &S) {
 341   return OS << "(" << S.S << ", " << (S.I ? std::to_string(*S.I) : "None")
 342             << ", " << S.B << ")";
 343 }
 344 bool fromJSON(const Value &E, CustomStruct &R) {
 345   ObjectMapper O(E);
 346   if (!O || !O.map("str", R.S) || !O.map("int", R.I))
 347     return false;
 348   O.map("bool", R.B);
 349   return true;
 350 }
 351
 352 TEST(JSONTest, Deserialize) {
 353   std::map<std::string, std::vector<CustomStruct>> R;
 354   CustomStruct ExpectedStruct = {"foo", 42, true};
 355   std::map<std::string, std::vector<CustomStruct>> Expected;
 356   Value J = Object{
 357       {"foo",
 358        Array{
 359            Object{
 360                {"str", "foo"},
 361                {"int", 42},
 362                {"bool", true},
 363                {"unknown", "ignored"},
 364            },
 365            Object{{"str", "bar"}},
 366            Object{
 367                {"str", "baz"}, {"bool", "string"}, // OK, deserialize ignores.
 368            },
 369        }}};
 370   Expected["foo"] = {
 371       CustomStruct("foo", 42, true),
 372       CustomStruct("bar", llvm::None, false),
 373       CustomStruct("baz", llvm::None, false),
 374   };
 375   ASSERT_TRUE(fromJSON(J, R));
 376   EXPECT_EQ(R, Expected);
 377
 378   CustomStruct V;
 379   EXPECT_FALSE(fromJSON(nullptr, V)) << "Not an object " << V;
 380   EXPECT_FALSE(fromJSON(Object{}, V)) << "Missing required field " << V;
 381   EXPECT_FALSE(fromJSON(Object{{"str", 1}}, V)) << "Wrong type " << V;
 382   // Optional<T> must parse as the correct type if present.
 383   EXPECT_FALSE(fromJSON(Object{{"str", 1}, {"int", "string"}}, V))
 384       << "Wrong type for Optional<T> " << V;
 385 }
 386
 387 TEST(JSONTest, Stream) {
 388   auto StreamStuff = [](unsigned Indent) {
 389     std::string S;
 390     llvm::raw_string_ostream OS(S);
 391     OStream J(OS, Indent);
 392     J.object([&] {
 393       J.attributeArray("foo", [&] {
 394         J.value(nullptr);
 395         J.value(42.5);
 396         J.arrayBegin();
 397         J.value(43);
 398         J.arrayEnd();
 399       });
 400       J.attributeBegin("bar");
 401       J.objectBegin();
 402       J.objectEnd();
 403       J.attributeEnd();
 404       J.attribute("baz", "xyz");
 405     });
 406     return OS.str();
 407   };
 408
 409   const char *Plain = R"({"foo":[null,42.5,[43]],"bar":{},"baz":"xyz"})";
 410   EXPECT_EQ(Plain, StreamStuff(0));
 411   const char *Pretty = R"({
 412   "foo": [
 413     null,
 414     42.5,
 415     [
 416       43
 417     ]
 418   ],
 419   "bar": {},
 420   "baz": "xyz"
 421 })";
 422   EXPECT_EQ(Pretty, StreamStuff(2));
 423 }
 424
 425 } // namespace
 426 } // namespace json
 427 } // namespace llvm